Skip to main content
Use these patterns when an integration needs linked proofs, version history, or tag-based lookup. The proof ID in API responses is qHash.

Typed reference: reference.type: "qhash"

The ownership-basic check accepts a reference object with type: "qhash" and id set to another proof’s ID. That creates a typed link from one proof to another.

Version chain: supersedes:<qHash> tag

When you update a skill or tool, or append to a session receipt, Proofable tags the new proof with supersedes:<previousQHash>. Newer proofs point back to older ones.
Call proofable_proofs_get with a qHash and traverse: "supersedes" to walk the chain.

Tag prefixes for discovery

A shared tag prefix such as topic: groups related proofs. Filter by prefix with tagPrefixesAll in proofable_proofs_get. To attach an agent profile to a project, see Connect agent context.

Large content

Inline proof content is limited to 50,000 characters. For larger artifacts, store the file elsewhere and attach a short proof that references it by URL or content hash.

Next

Proofs

How proofs work on Proofable.

Guard an agent action

Check permissions before a tool runs.

ownership-basic

Check inputs for content and references.
Last modified on October 25, 2026